Industrial Process Control with Human Override Prediction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Human operators frequently intervene in industrial processes controlled by automated systems, leading to suboptimal operation due to perceived malfunctions or unexpected failures, causing processes to deviate from optimal trajectories defined by optimality criteria.
Innovation Solution
A method using a process controller that determines control outputs based on state variables and queries a trained machine-learning model to predict the propensity of human override, modifying these outputs or constraints to reduce intervention likelihood, and conveying messages to enhance transparency and understanding for human operators.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If the automated control system operates strictly according to optimality criteria, then process efficiency and resource utilization are improved, but human operators perceive the control actions as implausible and intervene frequently
Solution Approach 1:
A machine learning model acts as an intermediary between the automated control system and human operators. This model predicts operator propensity to override and provides explanations for control actions, making the system's decisions more transparent and acceptable to operators without sacrificing optimality
Solution Approach 2:
The system implements feedback by using the machine learning model to predict operator reactions to control actions. This feedback loop allows the system to adjust its communication strategy and provide explanations that increase operator trust and reduce unnecessary interventions
2Speed
If the process controller makes aggressive control moves to reach target values quickly, then response time and productivity are improved, but operators perceive these moves as potentially unsafe and take manual control
Solution Approach 1:
The machine learning model serves as an intermediary that translates aggressive control actions into operator-friendly explanations. It predicts how operators will perceive these actions and provides contextual information that builds trust while maintaining fast response times
Solution Approach 2:
The system changes the parameter of operator perception by providing explanatory information alongside control actions. This transforms the operator's understanding of aggressive control moves from potentially unsafe to necessary and well-reasoned
Data Source
AI summary
A method for controlling an industrial process includes: determining, by a process controller, based at least in part on a set of current values and/or past values of state variables of the industrial process, a set of control outputs to be applied to at least one actor and/or lower-level controller configured to cause a performing of at least one physical action on the process; querying, based on at least a subset of the set of current values and/or past values of state variables and on at least a subset of the set of control outputs, a trained machine-learning model configured to output a classification value, and/or a regression value, that is indicative of a propensity of a watching human operator to at least partially override the control outputs delivered by the process controller; and determining that the classification value, the regression value, and/or the propensity, meets a predetermined criterion.